How to try out Sora, OpenAI’s new AI video generation platform

OpenAI is back to shake things up in the generative AI space with Sora, its new generative AI platform which can create videos from text prompts.

Sora can generate minute-long videos that are almost lifelike, with multiple shots and subjects, detailed objects and background elements, and even motion patterns. What’s more, you can even feed the platform with a still image, and generate a video based on the content of the video via textual prompts. The platform is being called a game changer for short-format video production, and hopefully, long-format video production in the near future as well.

Having gone through what the platform can do though, the next obvious question that arises is how to use it. If you’re looking for the answer to the same, read on to find out.

How can you try out Sora?

If you’re looking to try out Sora, unfortunately, you may have to wait for a bit longer. Just how much longer remains unknown as of now too, since OpenAI hasn’t mentioned any potential launch dates for the platform just yet.

ALSO READ: Sora is a generative AI platform that creates great-looking videos from just text

Sora is currently undergoing testing by a select group of content creators and filmmakers, to iron a couple of niggles that the platform seems to have. For starters, it still somewhat struggles to simulate physics in the real world.

Secondly, there have also been instances wherein it’s struggled to understand cause and effect. In one example cited by OpenAI, a man is shown eating a cookie generated on the platform. However, the cookie, afterward, appears to have no bite marks on it.

The platform is expected to be rolled out when these issues are ironed out. Now, it could be rolled out as a standalone web interface, similar to the likes of ChatGPT and Dall-E, or even as an API that developers can integrate into their own applications, presumably for video editing. Whether the platform will be accessible only on laptops or desktops, or even smartphones though, also remains unknown as of now.

Other AI-enabled video editing platforms to try while you wait for Sora

Google recently announced  a platform similar to Sora, called VideoPoet. In essence, VideoPoet is a large language model (LLM) that can produce videos through text-based prompts. The platform can also make edits to your videos based on the feedback you provide it with – such as trimming them, erasing objects and audio, or even cropping the frame.

Google claims that it has used as many as a billion image-text pairs and 270 million video samples to train the LLM. However, the LLM is capable of producing videos only up to eight-to-ten seconds long as of now. Moreover, similar to Sora, the LLM is also currently undergoing testing, and a launch date for the same remains unknown as of now.

While you may have to wait just a bit more to try platforms such as Sora and VideoPoet though, you can always try to use platforms such as Synthesia. Synthesia is a platform that allows you to create videos with human presenters via text-based prompts, wherein you can choose over 160 AI-based avatars, who can speak in over 140 languages.

There are plenty of other AI-enabled video creation platforms too, such as Runway, Unscreen, and Descript, which you can try your hand at before Sora is rolled out. Bear in mind, however, that they are for editing existing videos rather than creating them from scratch.

Nevertheless, these platforms come with several cool features, including the ability to remove backgrounds and objects in videos, generate captions, create animations, and much more.